A model that makes a clip with sound right away: speech, noises and background are created together with the picture. It animates one photo, builds one clip from several, and can lift the result to 2K and 4K.
Create a video with MiniMax H3
| What it does | Video from a photo and from text |
|---|---|
| Resolutions | 480p, 768p, 2K and 4K (upscaled from 768p) |
| Length | 5, 10 or 15 seconds (5 or 10 in “Video”) |
| Sound | Yes, always created |
| Several photos in a clip | Up to 9 |
| Last frame | Can be set |
| Price per clip | from $0.33 to $3.12 |
| Content filter | Yes |
Dialogue, noises and background are created together with the video and cannot be switched off. For ads this saves a separate voice-over.
The clip is first shot at 768p, and for 2K and 4K an upscale lifts it. The price is higher, but the final file is larger.
In “Photo → video” the length is 5, 10 or 15 seconds, in the “Video” tab 5 or 10.
The first five photos are included in the price, from the sixth there is a surcharge of about $0.10 each. Nine photos add about $0.42.
You can add a final frame to the source photo: the model joins them with movement.
H3 shoots at 480p and 768p and gets 2K and 4K by upscaling. H3 Max has 1080p as a separate quality level. Both models make a clip with sound.
No, sound is always created. If you do not need it, you can use the clip without it in your own editing.
The first five are included in the clip price. From the sixth to the ninth there is a surcharge of about $0.10 per photo, and the final price is shown before you run.
The model has a content filter and may reject such photos. For people, Seedance 1.5 in the “Photo → video” tab works more reliably.
No. You pay for each clip at the price shown before you run.